Another technique which can be used to analyze the data involves spatial cross-correlation techniques. First spatial anomalies on the parcel are visually identified and marked as patterns of interest. It is then investigated if these patterns of interest occur in other images. Spatial cross-correlation is robust in the sense that even when noise is present in the image, it is still able to determine if the patterns of interest are present in other images. This is a valuable property of this technique since noise in the images can severely influence the analysis and interpretation of the image. Until now, this technique is applied only to simulated data to get familiar with the technique and investigate the usefulness of it. The results look promising, therefore in future work this technique must be applied on obtained data and analyzed further. Currently, more data is being collected such that more extensive analyses can be carried out. Especially more work with regards to NDVI curves and their relationships with temporal variables needs to be done. In a next step the found relations must be formulated in an empirical model and combined with a crop growth and weather forecast model to cover the predictive aspect. Also the web application needs to be further developed to give more insight in the found relationships. Eventually this must lead to the creation of “Task Map 2.0” and “AgroGIS”.
5. Conclusion